paper-reviewlisted

# Paper Review Produce a review that helps the reader judge the paper: what it contributes, which claims the evidence supports, and which concerns could change the assessment. Be fair to strengths and weaknesses. Do not invent faults to fill a quota or assume that a paper deserves rejection. ## Working across assistants This skill uses ordinary Markdown and the host's available capabilities. It does not require a particular model, provider SDK, named search tool, or companion skill. See [runtime guidance](references/runtime-compatibility.md) when a capability is missing or the skill is embedded in an API workflow. Before reviewing: - Identify the supplied paper, any supplement, and the user's purpose: self-review, referee report, focused critique, or revision of an existing review. Respect that scope; a reference-only request does not require a new full review. - Read enough of the abstract and method to identify the topic before researching. Read the full paper, including tables and bibliography, before drawing conclusions. If only an abstract or excerpt is accessible, give a scoped critique and state what you could not assess. - Use the requested language and venue form. If neither is specified, use the user's language and the default structure below. Current user-provided review forms take precedence over bundled examples. - Check access to paper reading, external sources, and file output. Resolve `references/` paths relative to this `SKILL.md`, not the paper's workin